Message ID | 1431687246-31039-3-git-send-email-sebastian.hesselbarth@gmail.com |
---|---|
State | New |
Headers | show |
On Friday 15 May 2015 12:54:06 Sebastian Hesselbarth wrote: > this is Marvell Berlin SoC DT for v4.2 including: > - GPLv2/X11 dual-licensing of DT files, > - a note about currently unstable DT bindings as we struggle to get > DT bindings right in the first place with limited documentation available, > - Rework of System/Chip Controller binding documentation and nodes to > simple-mfd. > I'm a bit confused by this. Are you sure that it's ok to apply this branch by itself without introducing regressions or bisection problems? It's fine to have unstable bindings for a while as you are ramping up, but that requires a lot more care on your side regarding the order in which you send the patches. I have received three pull requests from you that are all based on -rc1, and from a brief look it seems like they are interdependent, which would not be ok. Arnd
On 15.05.2015 17:33, Arnd Bergmann wrote: > On Friday 15 May 2015 12:54:06 Sebastian Hesselbarth wrote: >> this is Marvell Berlin SoC DT for v4.2 including: >> - GPLv2/X11 dual-licensing of DT files, >> - a note about currently unstable DT bindings as we struggle to get >> DT bindings right in the first place with limited documentation available, >> - Rework of System/Chip Controller binding documentation and nodes to >> simple-mfd. >> > > I'm a bit confused by this. Are you sure that it's ok to apply this branch > by itself without introducing regressions or bisection problems? Ach, dammit. You are right, the patches reworking chip/sys ctrl nodes are currently un-bisectable. I do have Acked-by's from subsystem maintainers, so I guess I'll have to squash the DT changes with the driver changes. > It's fine to have unstable bindings for a while as you are ramping up, > but that requires a lot more care on your side regarding the order in > which you send the patches. I have received three pull requests from > you that are all based on -rc1, and from a brief look it seems like > they are interdependent, which would not be ok. Yes, the rework of the drivers and the corresponding DT changes depend on each other. Sorry for the confusion. Sebastian